Policy Transfer via Kinematic Domain Randomization and Adaptation

Ioannis Exarchos
Ioannis Exarchos
Yifeng Jiang
Yifeng Jiang
Wenhao Yu
Wenhao Yu
被引用0|引用|浏览18
其它链接arxiv.org

摘要

Transferring reinforcement learning policies trained in physics simulation to the real hardware remains a challenge, known as the "sim-to-real" gap. Domain randomization is a simple yet effective technique to address dynamics discrepancies across source and target domains, but its success generally depends on heuristics and trial-and-er...更多

代码

数据

您的评分 :
0

 

标签
评论